重庆理工大学学报2025，Vol.39Issue(1)：75-82,8.DOI:10.3969/j.issn.1674-8425(z).2025.01.010
融合项目评分不确定度的多属性深度神经协同推荐模型
Multi-criteria deep neural network collaborative filtering recommendation model incorporating item rating uncertainty
摘要
Abstract
Deep learning-based recommendation models have garnered considerable academic attention for their ability to explore complex interactions between users and items.However,the majority existing recommendation methods rely solely on users'single rating information during model training,neglecting users'preferences across different item criteria.This limitation compromises the accuracy of recommendations.To address it,we propose a multi-criteria deep neural network collaborative filtering recommendation model incorporating item rating uncertainty to capture users'multidimensional preference behavior.Meanwhile,to thoroughly consider the overall consistency of user preferences for item criteria,we introduce the item rating uncertainty to extract personalized criteria of items.Then,the item rating uncertainty is employed as weight factor for multi-criteria ratings to rectify the initial prediction results.By leveraging the adjusted multi-criteria ratings to predict user preferences,our model offers more accurate recommendations for users.Experimental results on two real-world datasets demonstrate the proposed model achieves a maximal increase of 4.3%in F1 score and 3.9%in NDCG score compared with those of sub-optimal method.These findings verify the improved recommendation efficacy and quality of our model.关键词
项目评分不确定度/多属性推荐模型/深度神经网络/协同过滤Key words
item rating uncertainty/multi-criteria recommendation models/deep neural networks/collaborative filtering分类
信息技术与安全科学引用本文复制引用
李昌兵,王霞,邓江洲..融合项目评分不确定度的多属性深度神经协同推荐模型[J].重庆理工大学学报,2025,39(1):75-82,8.基金项目
国家自然科学基金项目(62272077,72301050) （62272077,72301050）
重庆市教委科学技术研究项目(KJQN202300605) （KJQN202300605）
